Security audit
Security checks for vulnerabilities and agentic risk
The skill does not appear malicious, but its public identity and broad automatic activation do not match its actual AI Studio philosophy-graph workflow.
Review before installing. This should be renamed and scoped to the AI Studio philosophy-graph workflow, or rewritten to genuinely support USA business migration planning. Its implicit invocation and generic triggers should be narrowed before broad use.
The skill presents itself as a USA business migration planner, but the README describes an unrelated philosophy visualization workflow in AI Studio. This identity mismatch can mislead routing, review, and user trust decisions, and may allow a skill to be invoked in contexts where its real behavior has not been properly evaluated.
The documentation claims the skill satisfies its requirement, but the described requirement is unrelated to USA business migration planning. Such contradictions weaken operator understanding of what the skill will actually do and can conceal capability drift or unauthorized repurposing, increasing the chance of unsafe invocation.
The trigger keywords include extremely generic terms like 'general-help', 'create', 'studio', and 'hello', which can cause the skill to activate in many unrelated conversations. Overbroad activation increases the chance of unintended invocation, context hijacking, and accidental routing of user requests into a mismatched workflow.
The example trigger phrases are highly templated and vague, centering on broad requests for 'help' or a 'practical workflow' rather than a tightly scoped task. This makes it easier for ordinary user prompts to match unintentionally, leading to overreach by the skill and possible interference with safer or more appropriate agent behavior.
The skill is labeled as a USA business migration planner, but its manifest and body describe an unrelated AI Studio philosophy-graph task. This mismatch can cause the wrong skill to be invoked for unrelated user requests, leading to unsafe routing, misleading outputs, and potential policy bypass if downstream systems trust the name or metadata.
The manifest description says to use the skill for vague categories like general help and implementation support without clear boundaries. Ambiguous invocation criteria make accidental or excessive routing likely, which is dangerous in agent ecosystems because users may receive the wrong specialized workflow without realizing it.
The documentation contradicts itself about what the skill is supposed to do, which makes the operating boundary unclear. Contradictory instructions increase the chance of incorrect activation and unreliable behavior, especially in agent systems that use documentation text to select tools or shape responses.
The trigger list includes extremely broad terms like 'general-help', 'create', 'studio', and 'hello', which can match ordinary conversation and unintentionally activate the skill. Overbroad activation expands the skill's reach beyond its intended scope and can hijack unrelated user sessions with irrelevant or misleading behavior.
The skill’s declared identity ('USA Business Migration Planner') does not match its actual body, which is about creating a philosophy graph in AI Studio. This kind of semantic mismatch can cause the wrong skill to activate or be trusted under false pretenses, leading to misrouting, unsafe delegation, and user confusion that undermines review and policy controls.
The manifest description includes extremely broad trigger terms such as 'general-help' and 'hello', which overlap with ordinary conversation. This can cause accidental invocation in unrelated sessions, expanding the skill’s effective scope and enabling inappropriate context capture or interference with safer, better-matched skills.
The trigger section lists vague standalone keywords without context, scope boundaries, or exclusions. Broad unauthenticated activation logic increases the chance of misfires and unauthorized use of the skill in conversations where the user never intended this workflow.
The default prompt uses broad, generic phrasing such as 'help me' around an unrelated task, which can cause accidental or overly permissive invocation. In systems that rely on prompt text for tool selection, this increases the chance of unintended activation and misrouting, potentially exposing users to irrelevant or unsafe behavior.
The trigger guidance includes extremely generic terms such as "hello" and broad categories like "general-help" and "create," which can cause the skill to activate for many unrelated user requests. Overbroad activation can route benign conversations into an irrelevant or misleading workflow, increasing the chance of unintended behavior and prompt-scope confusion.
The trigger keywords include extremely broad terms like 'hello', 'create', and 'general-help', which can cause the skill to activate for unrelated user requests. In combination with the skill’s already-confused identity, this increases the risk of misrouting users into an inappropriate or unreviewed workflow.
The trigger sentences use common, open-ended phrasing that could match ordinary user requests without clear boundaries. Because the skill’s documented purpose is inconsistent, these loose activators make accidental execution more dangerous by broadening exposure beyond users explicitly seeking this exact capability.
Natural-language content throughout the README alternates between English and Chinese, but it does not state whether the skill will respond in a user-selected language or explain a justified locale restriction. This can violate language/locale policy expectations if users are implicitly forced into a mixed-language experience without opt-in.
The core description includes substantial Chinese text embedded in otherwise English instructions, but the file does not state whether responses should follow the user's preferred language or why this locale mix is required. This can create an implicit language constraint without user opt-in.
The file is written entirely as a Chinese-locale skill variant and does not indicate any user choice or opt-in for language preference. Because the instructions and activation behavior are presented only in this locale, the skill may impose a language constraint without documenting user selection or justification.
The title claims one function while the surrounding sections describe a different task and workflow. Internal contradictions weaken operator and user understanding, making it easier for an unrelated or improperly scoped skill to be invoked in sensitive contexts without clear informed intent.
The skill is branded as a USA business migration planner, but its description and prompt refer to an unrelated task about creating a philosophy visualization in Chinese. This mismatch can mislead routing systems and users, causing the skill to be invoked in the wrong contexts and making it harder to assess what actions or outputs are actually intended.
The default prompt mixes English and Chinese without any indication of user preference, which can force an unexpected language or locale context during invocation. While not as severe as direct code execution risks, it can confuse users, degrade consent around language selection, and increase the chance of incorrect task routing when paired with already inconsistent metadata.
Enabling implicit invocation without tightly scoped triggers allows the skill to activate based on vague similarity rather than clear user intent. Given the metadata mismatch elsewhere in this file, unintended invocation becomes more dangerous because the skill may be selected when users are asking for something entirely different.
The requirement statement embeds Chinese text inside an otherwise English document, and the usage signals repeat that mixed-language phrasing without any note about language selection or locale scope. This may effectively impose a language expectation without user opt-in or justification.
The activation scope is underspecified: the skill lists keywords and trigger sentences, but does not define clear constraints, boundaries, or disqualifying cases. This ambiguity makes accidental invocation more likely and can cause the agent to apply the skill outside its intended context, reducing reliability and creating unsafe prompt-routing behavior.
The main English README embeds substantial Chinese-language content in the requirement description without indicating that the user chose this language or that the skill is region-specific. This can impose a locale/language assumption on users and conflicts with the policy requiring choice or documented justification.
No suspicious patterns detected.
